The present invention relates to computer capture of object motion. More specifically, embodiments of the present invention relate to capturing of facial movement or performance of an actor.
Traditional computer animation of objects is determined by users known as animators. These animators are skilled artists who would specify movement of objects, such as people, within a computer environment. As a model for human movement, many animators often referred to how they moved, through the use of mirrors, video cameras, or the like.
Animation has also been based more directly upon physical movement of actors. This animation is known in the film industry as motion-capture or performance capture. In such cases, an actor is equipped with a suit with a number of markers, and as the actor moves, a number of cameras track the positions of the markers in space. This technique allows the actor's movements and expressions to be captured, and the movements and expressions can then be manipulated in a digital environment to produce whatever animation is desired.
One difficulty with prior motion capture techniques is that they often fail to produce high quality results with respect to capturing facial motion. Facial motion is very detailed and capturing the fine movements of an actor, or failing to do so, has a significant impact on the end product. Simply scaling current techniques to capture more markers is not practical as the data management, storage and processing capabilities are barely able to handle current marker density. High fidelity motion-capture, particularly video based data of the face, generates large volumes of digital data. Further, the fine-grain motion of the face is often lost in the noise inherent in stationary camera motion capture systems.
One solution to this problem has been to use separate and non-simultaneous face and body motion capture. A drawback to this approach is that it requires substantial duplication of effort by the actor and the crew, as each scene must be performed and captured at least twice. Another difficulty arises in that the actor's second performance may not correspond closely enough with the first, which affects the appearance and quality of the end result by making the correlation of the data from the multiple performances difficult.
Another solution is a simultaneous face and body motion capture using fixed position cameras and/or movable platform mounted cameras 100, such as is shown in FIG. 1. Capturing detailed facial motion generally involves tracking a large number of markers 140 placed on the actors' faces throughout a fixed capture volume 120, which is defined by the stationary motion capture cameras. In addition to the facial markers 140, markers 130 are placed on the actors' bodies, averaging a total of about 250 marker points per actor. For a scene with several actors, the total number of markers may be well over a thousand.
This abundance of markers creates a correspondingly large amount of data to be processed. It is sometimes difficult to accurately identify the markers associated with each actor 110 in a scene and obtaining sufficient resolution of the closely spaced facial markers 140 presents further complexities. In addition, because data from both the face and the body is captured together, it is necessary to process all of the data in order to determine whether sufficient facial motion data was recorded. Similarly, feedback cannot be given to the director or actor regarding the overall movement in the scene until all of the data has been processed. Waiting for the combined facial and body data to be processed significantly increases the delay between the initial capture and any reshoots that are necessary, likely causing production and scheduling problems, as well as increasing costs.
Prior head-mounted cameras have had several obstacles, such as interfering with the performance of the actor, either due to the discomfort of wearing the device or from the mere presence of the device in front of the actor's eyes or mouth, or failing to capture images of an adequate portion of the face for quality reconstruction. In addition, prior head-mounted cameras experience difficulty in maintaining position or in repositioning the camera.
Accordingly, an improved system for capturing and processing facial motions of an actor that increases accuracy while minimizing processing time and difficulty is desired. In addition, it would be desirable to have a head-mounted camera that is unobtrusive to the actor, but is capable of imaging a substantial portion of the face, and is easily repositioned.